Alone in the Dark (1992 DOS)

NHG

Edit history:

NHG: 2016-05-30 10:42:31 am

2016-05-29 01:27:43 pm

Shut up and go back to another earth

Fast intervention: Tigrou, we can actually die from falling in E5R10 and E5R11, the only spot, as far as I know, where there are 4 different heights ^^
Thanks for the autorization Lot', I do that tonight I think. See ya =p

Edit: I put times that into the board. Do I have to put some other times ? Right now I can't think of other alternative paths who may worth it. Maybe the key grabbing ? If I do, it'll be a "with viewer" chrono (wich is not the case of those that I wrote into the board). About that, I also timed the "Jump trick path" for Pregzt with the viewer. I optimized it as hard as I could: I pressed turn just before the climb in E6R0 so she instant-turns back once she is up to the cliff, straight line to reach E6R5 and I was close enough E6R6 to reach it without reactivating the jump-trick. Even like that it does not worth it anymore. A viewer performance is ~5sec behind the good-old clip path in E6R0 without viewer (45sec against 50sec). I did not put that into the board as it is not really an alternative path asking changings in the whole run, it is just another path as itself . But I can if you want !

Fast intervention: Tigrou, we can actually die from falling in E5R10 and E5R11, the only spot, as far as I know, where there are 4 different heights ^^
Thanks for the autorization Lot', I do that tonight I think. See ya =p

Edit: I put times that into the board. Do I have to put some other times ? Right now I can't think of other alternative paths who may worth it. Maybe the key grabbing ? If I do, it'll be a "with viewer" chrono (wich is not the case of those that I wrote into the board). About that, I also timed the "Jump trick path" for Pregzt with the viewer. I optimized it as hard as I could: I pressed turn just before the climb in E6R0 so she instant-turns back once she is up to the cliff, straight line to reach E6R5 and I was close enough E6R6 to reach it without reactivating the jump-trick. Even like that it does not worth it anymore. A viewer performance is ~5sec behind the good-old clip path in E6R0 without viewer (45sec against 50sec). I did not put that into the board as it is not really an alternative path asking changings in the whole run, it is just another path as itself . But I can if you want !

LotBlind

2016-05-30 01:40:24 pm

tigrou:

You now have AitD 2 and 3 as you noticed... I know someone's going to ask about that at some point and anyone who wants to run 2 or 3 will probably get SOMETHING out of being able to work with the viewer even if there's no OOB.

I'm looking at some scripts now and will throw questions your way in the private thread later.

NHG: I'm looking at the game scripts which will take a long time to go through (there's over 500), but it's really fascinating actually. I'll be communicating with NHG in a private thread about it because I'll have to ask him questions about it. If you like, ask him to include you in the thread but if I find anything worth mentioning I'll be sure to bring it up here.

You can definitely include even some more obscure ideas in the table such as doing the E2 key detour and various ways of doing the same area, just to make it clear it has been tested, and also for possible in-bounds runs and such. I think you should always use the viewer right now and make notes about stuff that's difficult to do without it and an estimate of how much slower safe strats will be. Or just have there be two rows, one for viewer and one for SS strats. Simple.

It's been brought up before - someone could do an in-bounds category for the game. It might look pretty similar to the old any% before clipping had been found. IMO you shouldn't be allowed to use clipping of any sort in the in-bounds, even if it doesn't really take you into the OOB as we understand that term. If you get bored waiting for me to finish with the scripts, maybe you want to see what kinds of changes are possible in the old run to make it a better example of this.

If you think it would make a better run if you DID allow clipping where it doesn't go OOB, you can argue for that. Then you could e.g. use chairs to get through locked doors in case that was faster somewhere.

Another thing I really wanna do is to create an "alternative" route just to showcase the different variations we discovered. It doesn't have to be super-optimized, but it should include as many obscure and strange solutions to everything as possible with no regard as to what's fastest. That's another idea you could start drafting. The upside of this is we could definitely show off using the chair and things like that. I think you'll agree there's no point in trying to do this in real time Tongue

- but that means something like the arrow grab might be doable.

Good luck whatever you choose!

tigrou:

You now have AitD 2 and 3 as you noticed... I know someone's going to ask about that at some point and anyone who wants to run 2 or 3 will probably get SOMETHING out of being able to work with the viewer even if there's no OOB.

I'm looking at some scripts now and will throw questions your way in the private thread later.

NHG: I'm looking at the game scripts which will take a long time to go through (there's over 500), but it's really fascinating actually. I'll be communicating with NHG in a private thread about it because I'll have to ask him questions about it. If you like, ask him to include you in the thread but if I find anything worth mentioning I'll be sure to bring it up here.

You can definitely include even some more obscure ideas in the table such as doing the E2 key detour and various ways of doing the same area, just to make it clear it has been tested, and also for possible in-bounds runs and such. I think you should always use the viewer right now and make notes about stuff that's difficult to do without it and an estimate of how much slower safe strats will be. Or just have there be two rows, one for viewer and one for SS strats. Simple.

It's been brought up before - someone could do an in-bounds category for the game. It might look pretty similar to the old any% before clipping had been found. IMO you shouldn't be allowed to use clipping of any sort in the in-bounds, even if it doesn't really take you into the OOB as we understand that term. If you get bored waiting for me to finish with the scripts, maybe you want to see what kinds of changes are possible in the old run to make it a better example of this.

If you think it would make a better run if you DID allow clipping where it doesn't go OOB, you can argue for that. Then you could e.g. use chairs to get through locked doors in case that was faster somewhere.

Another thing I really wanna do is to create an "alternative" route just to showcase the different variations we discovered. It doesn't have to be super-optimized, but it should include as many obscure and strange solutions to everything as possible with no regard as to what's fastest. That's another idea you could start drafting. The upside of this is we could definitely show off using the chair and things like that. I think you'll agree there's no point in trying to do this in real time :P - but that means something like the arrow grab might be doable.

Good luck whatever you choose!

NHG

Edit history:

NHG: 2016-05-30 03:28:38 pm

2016-05-30 03:27:36 pm

Shut up and go back to another earth

AITD2/3: Actually, at the very beginning I planned to run all of the 3 episodes. I eventually lost motivation about the last episode as there was almost no trick at all, as opposed to the 2nd episode wich was really interesting to run. All that said, I still want to do it, besides a speedrun showed up on Youtube and it is... Well it is not bad but it more a fast walkthrough, the guy clearly did not have "speedrunner mecanisms" in mind when he did it. He kills almost every ennemy wich is not necessary at all, he does not optimize moves/menuing and there isn't a single trick (I said there were not many, not that there was not). Anyway maybe viewer will help me to find tricks and motivate myself to the work. I'll be more upset about re-doing the 2nd one as it took me a lot of practice but... Meeeh, if it has to be done, I'll do it too =p

Lot': I'll add some informations and times into the board but I don't know if I'll time all possible variations. There are now many and.. Well basically I just wanted to run this dayum game T_T I'll give an effort to add some sequences tomorrow, those susceptible to be interesting (E1 OOB, Key grabbing etc...)

I share the same opinion about the in-bounds category, it'll make sense if it's a "no OOB/clip" run. It would be close the any% indeed, something like my old 10:56 but with some variations though. I'll write the path it shouldn't be long.

About the other possible routes, it could be interesting as there are many possibilities (low% with manimum items) but I won't have the patience for that, I'm more an any% guy and god damn it it's been 3years, I wanna move on t_t

I'll give some shots very soon but I can already tell it will be, and by far, my shortest speedrun =p

AITD2/3: Actually, at the very beginning I planned to run all of the 3 episodes. I eventually lost motivation about the last episode as there was almost no trick at all, as opposed to the 2nd episode wich was really interesting to run. All that said, I still want to do it, besides a speedrun showed up on Youtube and it is... Well it is not bad but it more a fast walkthrough, the guy clearly did not have "speedrunner mecanisms" in mind when he did it. He kills almost every ennemy wich is not necessary at all, he does not optimize moves/menuing and there isn't a single trick (I said there were not many, not that there was not). Anyway maybe viewer will help me to find tricks and motivate myself to the work. I'll be more upset about re-doing the 2nd one as it took me a lot of practice but... Meeeh, if it has to be done, I'll do it too =p

Lot': I'll add some informations and times into the board but I don't know if I'll time all possible variations. There are now many and.. Well basically I just wanted to run this dayum game T_T I'll give an effort to add some sequences tomorrow, those susceptible to be interesting (E1 OOB, Key grabbing etc...)

I share the same opinion about the in-bounds category, it'll make sense if it's a "no OOB/clip" run. It would be close the any% indeed, something like my old 10:56 but with some variations though. I'll write the path it shouldn't be long.

About the other possible routes, it could be interesting as there are many possibilities (low% with manimum items) but I won't have the patience for that, I'm more an any% guy and god damn it it's been 3years, I wanna move on t_t

I'll give some shots very soon but I can already tell it will be, and by far, my shortest speedrun =p

tigrou

Edit history:

tigrou: 2016-05-31 03:13:23 am

tigrou: 2016-05-31 03:13:13 am

tigrou: 2016-05-31 03:13:02 am

tigrou: 2016-05-31 03:12:40 am

tigrou: 2016-05-31 01:48:05 am

tigrou: 2016-05-31 01:24:22 am

2016-05-31 12:33:59 am

@LotBlind : I am personally not a fan of AITD2 or AITD3 but I can bring support for it in the room viewer if its not too hard. Maybe it already work out of the box. What will probably not work is the "link to DOSBOX" feature. I will check when I will get some free time. Btw, it only make sense to do this if there is at least one guy interested of doing speedrun on it.

NHG

Edit history:

NHG: 2016-05-31 05:42:07 am

2016-05-31 05:27:29 am

Shut up and go back to another earth

Of course you're not a fan, who could be ? Those sequels are awful x) but if I can complete the trilogy I'll be glad ! But don't worry you can take your time, it won't bother me to play normally for a while x)

I added some infos into the board. I think there's everything about the two major paths.

About the any No-clip%, as I already said it wont be so much different than the 10:56 speedrun. If I don't forget anything...

E0: Grab lantern + cover, go down
E1: Grab oil, use it, grab key, use it on chest, take saber, go down using zomby-chicken trick
E2: Grab matches, grab first aid case, use cover on painting (can optimize that by using it far from the wall), open first aid case, use flask in R7 to cancel arrow, grab fakebook, grab study key, use matches, use fakebook, grab talisman, go down
E3: Grab pot of soup, grab chest key, use pot of soup, grab lighter (or not ?), use study key, use saber, go down
E5/6: No change at all here except the use of the jump-trick, wich also may be useful in E5R10 to go faster (also may you fly over the first maze part). Exit by study path is now possible and saves considerable time
E3: Leave the house, live your life !

The time should be something around 10 minuts

Of course you're not a fan, who could be ? Those sequels are awful x) but if I can complete the trilogy I'll be glad ! But don't worry you can take your time, it won't bother me to play normally for a while x)

I added some infos into the board. I think there's everything about the two major paths.

About the any No-clip%, as I already said it wont be so much different than the 10:56 speedrun. If I don't forget anything...

E0: Grab lantern + cover, go down
E1: Grab oil, use it, grab key, use it on chest, take saber, go down using zomby-chicken trick
E2: Grab matches, grab first aid case, use cover on painting (can optimize that by using it far from the wall), open first aid case, use flask in R7 to cancel arrow, grab fakebook, grab study key, use matches, use fakebook, grab talisman, go down
E3: Grab pot of soup, grab chest key, use pot of soup, grab lighter (or not ?), use study key, use saber, go down
E5/6: No change at all here except the use of the jump-trick, wich also may be useful in E5R10 to go faster (also may you fly over the first maze part). Exit by study path is now possible and saves considerable time
E3: Leave the house, live your life !

The time should be something around 10 minuts

tigrou

Edit history:

tigrou: 2016-06-14 02:39:20 am

tigrou: 2016-06-04 05:43:18 pm

tigrou: 2016-06-04 01:56:09 pm

tigrou: 2016-06-04 10:03:25 am

tigrou: 2016-06-04 10:03:10 am

tigrou: 2016-06-04 09:11:31 am

tigrou: 2016-06-04 09:11:07 am

2016-06-04 09:08:48 am

Hi there!

I managed to add support for AITD2/3 games.
Everything should work : rooms, link to dosbox feature and model viewer.
It was not that hard because all games use the same engine. Btw, there was differences which required small changes.

Here is more details :

Roomviewer
If you try to open AITD2/3 rooms with 1.0.4 version it won't work.
This is because the number of floors and rooms is hardcoded in the room viewer (for simplicity).
I have modified the room viewer so that the number of rooms and floor is calculated from the data itself.

In the process, I discovered a new room in AITD1 : E7R2. Before you start being excited by this, this is just a dummy room which contains one collider.
It is used for the final scene (with the taxi guy).

About AITD2/3 rooms :
=> it seems to use tons of colliders and triggers. Could this explain why AITD2/3 are slower than original game ?
=> AITD2 : the ship rooms are interesting. anyway it shows you one limitation of AITD engine : it cannot handle non-orthogonal walls.
=> on some floors, you will notify there is lot of small cubic rooms put aside, in a row, with one collider.
I don't know what it is used for. At first, I thought it was used for showing cutscenes with different camera angles (like E7R2) but there isn't any camera data associated.

DOSBOX
To make it work, you should proceed like in AITD1 : when player is at initial position (after intro), press "L" and wait for the magic to happen.

Modelviewer
It worked out of the box without any changes. Anyway colors are wrong.
This is because games use different color palette. After manually extracting them from games and importing them in the viewer it works flawlessly. It seems AITD2/3 borrows a lot of models from AITD1 (they only changed colors or adapted it a little bit).

I tested this but not fully (eg : by finishing AITD2/3). I did not go further than the first rooms. So there might be some bugs later in the story.

Hi there!

I managed to add support for AITD2/3 games. 
Everything should work : rooms, link to dosbox feature and model viewer.
It was not that hard because all games use the same engine. Btw, there was differences which required small changes.

Here is more details :

[b]Roomviewer[/b]
If you try to open AITD2/3 rooms with 1.0.4 version it won't work.
This is because the number of floors and rooms is hardcoded in the room viewer (for simplicity).
I have modified the room viewer so that the number of rooms and floor is calculated from the data itself.

In the process, I discovered a new room in AITD1 : E7R2. Before you start being excited by this, this is just a dummy room which contains one collider.
It is used for the final scene (with the taxi guy).

About AITD2/3 rooms :
=> it seems to use tons of colliders and triggers. Could this explain why AITD2/3 are slower than original game ?
=> AITD2 : the ship rooms are interesting. anyway it shows you one limitation of AITD engine : it cannot handle non-orthogonal walls.
=> on some floors, you will notify there is lot of small cubic rooms put aside, in a row, with one collider. 
I don't know what it is used for. At first, I thought it was used for showing cutscenes with different camera angles (like E7R2) but there isn't any camera data associated.

[b]DOSBOX[/b]
To make it work, you should proceed like in AITD1 : when player is at initial position (after intro), press "L" and wait for the magic to happen.

[b]Modelviewer[/b]
It worked out of the box without any changes. Anyway colors are wrong.
This is because games use different [url=https://i.imgur.com/GDcv2QD.png]color palette[/url]. After manually extracting them from games and importing them in the viewer it works flawlessly. It seems AITD2/3 borrows a lot of models from AITD1 (they only changed colors or adapted it a little bit).

I tested this but not fully (eg : by finishing AITD2/3). I did not go further than the first rooms. So there might be some bugs later in the story.

LotBlind

2016-06-04 10:58:41 am

NHG: You could just get the cellar key and do the old worm skip instead (for in-bounds)? Or isn't the sabre + key detours going to take too much time?

Tigrou: Ha! I knew it... You're going to be very popular in the AitD wiki forums as well as with speedrunners here. At least if it helps someone find something truly interesting in 2 and 3. Tongue

Fingers crossed!

I wish my wrist wasn't hurting like this, I'm trying to look into the AitD scripts more tonight.

NHG

2016-06-04 11:36:19 am

Shut up and go back to another earth

Oh gosh I said take your time xD Well fine, good job anyway ^^ I'll try that soon on the last episode first as I did not run it at all. I re-watched the current speedrun yesterday and... Yeah, definitely I want the WR to be better. Actual time is something like 40:00 but he looses so much time killing absolutely everyone (wich means he takes useless weapons/munitions/flasks all the time, wich also means lot of menuing), and he does not even skip all the cutscenes. I think even without finding any shortcut (and I hope we will), it'll be easy enough to save something like 5-6 minuts.

LotBlind: I think I kept the classic path because before clip era, Chtonian path including cellar key did not worth it (well I remember I was really disapointed at first because I found this path was cool but it wasn't faster for a time). Let me think... Saber 17sec, for the study key let's say 5sec BUT no boulder -11 sec and study room exit -21 sec. Into the board, we can see that a study path is 29 sec while a cellar path is 28 so almost no difference, in both cases the added detour'll ask to grab an extra item. The key is further and in both cases there are 3 more doors to open. Theorically there's no way for the Chtonian path to be better.

I'm not going to keep it but yesterday i recorded a "not so bad" run, time was a lot shorter than I thought =p I'll keep the surprise but sub 7min by far !

Oh gosh I said take your time xD Well fine, good job anyway ^^ I'll try that soon on the last episode first as I did not run it at all. I re-watched the current speedrun yesterday and... Yeah, definitely I want the WR to be better. Actual time is something like 40:00 but he looses so much time killing absolutely everyone (wich means he takes useless weapons/munitions/flasks all the time, wich also means lot of menuing), and he does not even skip all the cutscenes. I think even without finding any shortcut (and I hope we will), it'll be easy enough to save something like 5-6 minuts.

LotBlind: I think I kept the classic path because before clip era, Chtonian path including cellar key did not worth it (well I remember I was really disapointed at first because I found this path was cool but it wasn't faster for a time). Let me think... Saber 17sec, for the study key let's say 5sec BUT no boulder -11 sec and study room exit -21 sec. Into the board, we can see that a study path is 29 sec while a cellar path is 28 so almost no difference, in both cases the added detour'll ask to grab an extra item. The key is further and in both cases there are 3 more doors to open. Theorically there's no way for the Chtonian path to be better.

I'm not going to keep it but yesterday i recorded a "not so bad" run, time was a lot shorter than I thought =p I'll keep the surprise but sub 7min by far !

NHG

Edit history:

NHG: 2016-06-05 10:49:12 am

NHG: 2016-06-04 11:39:28 am

NHG: 2016-06-04 11:38:40 am

NHG: 2016-06-04 11:37:38 am

2016-06-04 11:37:21 am

Shut up and go back to another earth

Oh, poopy I can't erase that... Sorry 'bout that !

Man, this last episode is soooo empty... Yet there are red hitboxes absolutely everywhere, it's insane. Until now I didn't find anything I already knew when I began to route it long time ago ( some tricky jump, a certain path avoiding trigger ennemies in a room...) The viewer didn't show me anything interesting but I'll give a closer look just to be sure. Also, when comes the Puma phase the viewer looses the link. No hurry, it's not a big deal but I'm just reporting that !

LotBlind

2016-06-05 12:56:46 pm

NHG: It's a good idea for you to be doing AitD 2 or 3 right now Tongue

Technically this thread is for AitD 1 and if you're going to make heavy notes for 2 or 3 you should probably put them on another thread just so people can find them easier. That or start editing the guide.

NHG

2016-06-05 02:32:03 pm

Shut up and go back to another earth

Ahah yeah yeah don't worry I'll do ! I just tried the viewer to see if it worked well and I could not stop myself to search things ^^ I'll finish with Alone 1 first anyway

LotBlind

2016-06-08 12:42:23 pm

Someone on the wiki wants to start some French language page on it... Here's the announcement if one of you wants to get involved.

tigrou

Edit history:

tigrou: 2016-06-09 06:28:51 am

tigrou: 2016-06-09 05:04:09 am

tigrou: 2016-06-09 02:37:59 am

tigrou: 2016-06-09 02:37:38 am

tigrou: 2016-06-09 02:18:52 am

tigrou: 2016-06-09 02:18:34 am

tigrou: 2016-06-09 02:18:27 am

2016-06-08 11:41:57 pm

Some interesting links about AITD that I found recently :

Alone in the Dark 3 - Making of Trailer
You can view how they created the rooms, the models and the scripts. This is for AITD3 but tools are probably the same as AITD1.

Alone in the dark : Epilogue: Call of Cthulhu
More info about tools etc...

Hardcore Gaming 101 : Alone in the Dark
A page with reviews of AITD series and related products.

Quote:

Someone on the wiki wants to start some French language

Maybe he will be interested by the high-res images the model viewer can produce (although they are not 100% accurate with the real thing). I have implemented support for gradient materials last week. Some models (like the mirror) looks lot better now.

NHG

Edit history:

NHG: 2016-06-14 05:36:21 am

NHG: 2016-06-13 09:38:14 am

NHG: 2016-06-09 06:30:01 pm

2016-06-09 03:30:36 pm

Shut up and go back to another earth

Gentleman, I am happy to say that the new World Record on Alone in the Dark 1 has a nice time of... 6:24 ! I'll upload the thing tomorrow, I passed the whole day on this, juggling between the game and Lovecraft's Dreamland as a break, and I need a serious breath of fresh air ! If want to be critical I would say that I failed a run start towards the end and that my last clip could be slightly faster, I indeed probably could have done 6:22 (2 minuts saved compared to the old run) but the run is satisfaying to me. All the things I wanted to realize perfectly are there (funny thing that one of my main difficulties is still to not touch this damn Gramophone), the clip through the library was so well-made that I had to wait a litlle at the exit to avoid being stuck by the Vagabond in the angle.

Anyway, I am happy ! I thank you ! I am motivated for Alone 3 ! And I need to rest ! See you tomorrow pals !

Edit: God damn it ! I just thought about another use of the jump-trick. I tested it and it saves ~4sec. So... Fuck me ! New improvment incoming...

Edit2: 6:21, that's it I quit !

Edit3:

I hope the sound'll be ok for the Quality test. I already said that somewhere but oddly enough since the computer works with Windows 10, the sound in Alone1 is not the same. The music is okay but some of the noises kinda spit, especially the footstep on tiles

Gentleman, I am happy to say that the new World Record on Alone in the Dark 1 has a nice time of... 6:24 ! I'll upload the thing tomorrow, I passed the whole day on this, juggling between the game and Lovecraft's Dreamland as a break, and I need a serious breath of fresh air ! If want to be critical I would say that I failed a run start towards the end and that my last clip could be slightly faster, I indeed  probably could have done 6:22 (2 minuts saved compared to the old run) but the run is satisfaying to me. All the things I wanted to realize perfectly are there (funny thing that one of my main difficulties is still to not touch this damn Gramophone), the clip through the library was so well-made that I had to wait a litlle at the exit to avoid being stuck by the Vagabond in the angle.

Anyway, I am happy ! I thank you ! I am motivated for Alone 3 ! And I need to rest ! See you tomorrow pals !

Edit: God damn it ! I just thought about another use of the jump-trick. I tested it and it saves ~4sec. So... Fuck me ! New improvment incoming...

Edit2: 6:21, that's it I quit !

Edit3: https://www.youtube.com/watch?v=HD2Dq75Kzro&feature=youtu.be

I hope the sound'll be ok for the Quality test. I already said that somewhere but oddly enough since the computer works with Windows 10, the sound in Alone1 is not the same. The music is okay but some of the noises kinda spit, especially the footstep on tiles

LotBlind

2016-06-15 12:21:45 pm

Whoops, I got no notification for this update...

Well well well... That's a run and a half! Can I request you don't submit anything until a bit later? I'll try to look into more code stuff as soon as I've got the energy for it, and if it looks like a segmented run can save time somewhere (aside from just optimizations and a few doors here and there) I still wanna do that.

A/V Music is a bit loud actually but not unsubmittable or anything. I'm probably going to turn it off for my run (less lag and I just don't like it). I can't hear the sfx problem though.

1:15 - can't you get a quick turn here? I don't really see a reason. If you are afraid of hitting the ghost, you can still get a quick turn without the run.
1:38 - That's weird, it usually changes the camera angle for me when I go to that corner. The fact that it didn't here is not bad unless it caused you to lose time but you obviously didn't really lose that much.
2:08 - so obviously getting the door clip here saves several seconds, less so for segmented but still: your library skip was indeed really quick.

Man that long OOB really looks ridiculous without helpers... Cheesy

5:05 - Tigrou: I wanted to ask you if you could do the maths for how to pass through that vertically rising door. Because you're waiting for it to open (and it can't be skipped like other doors even by s/l) you can reposition yourself anywhere you like. What's theoretically the best point to start the run and the best direction to move when you take into account that the next corner requires a slightly shorter rotation the further you're already rotated towards the left. Do you see what I mean?

5:20 - Wait why does that work? You do the jump trick BEFORE entering E5... it's supposed to reset isn't it? Why doesn't it reset?

How much time did you lose in the OOB and taking the weird corridor in E5 when coming back? I'm asking so I know if a segmented run can do a lot better there...

Whoops, I got no notification for this update...

Well well well... That's a run and a half! Can I request you don't submit anything until a bit later? I'll try to look into more code stuff as soon as I've got the energy for it, and if it looks like a segmented run can save time somewhere (aside from just optimizations and a few doors here and there) I still wanna do that.

A/V Music is a bit loud actually but not unsubmittable or anything. I'm probably going to turn it off for my run (less lag and I just don't like it). I can't hear the sfx problem though.

1:15 - can't you get a quick turn here? I don't really see a reason. If you are afraid of hitting the ghost, you can still get a quick turn without the run.
1:38 - That's weird, it usually changes the camera angle for me when I go to that corner. The fact that it didn't here is not bad unless it caused you to lose time but you obviously didn't really lose that much.
2:08 - so obviously getting the door clip here saves several seconds, less so for segmented but still: your library skip was indeed really quick.

Man that long OOB really looks ridiculous without helpers... :D

5:05 - Tigrou: I wanted to ask you if you could do the maths for how to pass through that vertically rising door. Because you're waiting for it to open (and it can't be skipped like other doors even by s/l) you can reposition yourself anywhere you like. What's theoretically the best point to start the run and the best direction to move when you take into account that the next corner requires a slightly shorter rotation the further you're already rotated towards the left. Do you see what I mean?

5:20 - Wait why does that work? You do the jump trick BEFORE entering E5... it's supposed to reset isn't it? Why doesn't it reset?

How much time did you lose in the OOB and taking the weird corridor in E5 when coming back? I'm asking so I know if a segmented run can do a lot better there...

NHG

2016-06-15 02:38:04 pm

Shut up and go back to another earth

I don't mind waiting before submitting it but what it has to do with your project for a segmented-run ? =o Is it for having both verified in the same time ?

I did not want to turn off sound or music even if I know it actually had an influence in the game. It would bother me to play the game without any sound.

1:15 => I think the fact that there are two items cancels the quick move between both of them because, theoretically I do the move. In any case it helps me to have the good position to come back with the run
1:38 => Oh ? I never had such a thing, thoses screens are the same as usual to me.
2:08 => Considering AFAWK it happens randomly and how difficult it is to have the "fast clip path" (when you can leave talisman room by the left), I didn't feel to reset until it happens. It is undeed a possible optimization

5:20 I have absolutely no idea why it does work despite the fact that it's technically another floor. I just tested activating it before leaving the maze, it worked and that was fine to me x) Maybe it only cancels when the character jumps or climb something ?

About the long OOB mmmh it's difficult to say how much time I lost. You can run more risky at 3:25 instead of walk back a little like I do. There's also the camera at 3:37 that I only trigger because I need to know where I am. A "viewer" performance can probably save ~5sec in the whole sequence. About the ghost corridor it is far less than that. My trajectory is probably unperfect but a perfect one would make 11 or 10 steps before reaching the wall (I make 12), I don't think you'll save a lot of time here.

I don't mind waiting before submitting it but what it has to do with your project for a segmented-run ? =o Is it for having both verified in the same time ?

I did not want to turn off sound or music even if I know it actually had an influence in the game. It would bother me to play the game without any sound.

1:15 => I think the fact that there are two items cancels the quick move between both of them because, theoretically I do the move. In any case it helps me to have the good position to come back with the run
1:38 => Oh ? I never had such a thing, thoses screens are the same as usual to me.
2:08 => Considering AFAWK it happens randomly and how difficult it is to have the "fast clip path" (when you can leave talisman room by the left), I didn't feel to reset until it happens. It is undeed a possible optimization

5:20 I have absolutely no idea why it does work despite the fact that it's technically another floor. I just tested activating it before leaving the maze, it worked and that was fine to me x) Maybe it only cancels when the character jumps or climb something ?

About the long OOB mmmh it's difficult to say how much time I lost. You can run more risky at 3:25 instead of walk back a little like I do. There's also the camera at 3:37 that I only trigger because I need to know where I am. A "viewer" performance can probably save ~5sec in the whole sequence. About the ghost corridor it is far less than that. My trajectory is probably unperfect but a perfect one would make 11 or 10 steps before reaching the wall (I make 12), I don't think you'll save a lot of time here.

Mr. Kelly R. Flewin

2016-06-15 07:28:33 pm

Learning to Stream

Sweet Zombie Jesus on a Pogo Stick! It's one thing to read about all these glitches, but to see that run put them into action is a head trip! Very nicely done!!

LotBlind

Edit history:

LotBlind: 2016-06-16 12:30:48 pm

2016-06-16 12:29:56 pm

NHG:

1:15 I see.
1:38 - whoops... I'm too used to using the viewer to do that so I couldn't remember what the actual view was like.
2:08 - It's NOT random and that stuff never is, it's your sub-pixels. If you hug the wall in a certain way (i.e. run next to it but slightly towards it) you'll greatly increase your chances of getting the clip. Use the viewer to see the PC's hitbox from up close and you can see the effect. It's probably not that difficult in the end, and will save time even if you have to set it up a little bit.

5:20 - Well maybe it's to do with the fact that between all the other floors, your position gets reset, whereas between 5 and 6, it depends where you hit the trigger at. However, why doesn't it give you the same effect coming from E5 to E6 then? The hover should be kept. EDIT: ... and so it is Smiley

You didn't put 2 and 2 together. Please don't go on another marathon rampage of attempts yet. Maybe we'll think of something more. Also why can't you just post a route video first so you can avoid this Tongue

Well you got lots of practice in!

I noticed if you do the jump-drop on one floor and land in the other one, it causes you to get stuck in place. Yay.

Mr. Kelly: it's the whip I'm holding

NHG:

1:15 I see.
1:38 - whoops... I'm too used to using the viewer to do that so I couldn't remember what the actual view was like.
2:08 - It's NOT random and that stuff never is, it's your sub-pixels. If you hug the wall in a certain way (i.e. run next to it but slightly towards it) you'll greatly increase your chances of getting the clip. Use the viewer to see the PC's hitbox from up close and you can see the effect. It's probably not that difficult in the end, and will save time even if you have to set it up a little bit.

5:20 - Well maybe it's to do with the fact that between all the other floors, your position gets reset, whereas between 5 and 6, it depends where you hit the trigger at. However, why doesn't it give you the same effect coming from E5 to E6 then? The hover should be kept. EDIT: ... and so it is :) You didn't put 2 and 2 together. Please don't go on another marathon rampage of attempts yet. Maybe we'll think of something more. Also why can't you just post a route video first so you can avoid this :P Well you got lots of practice in!

I noticed if you do the jump-drop on one floor and land in the other one, it causes you to get stuck in place. Yay.

Mr. Kelly: it's the whip I'm holding

NHG

2016-06-16 01:29:24 pm

Shut up and go back to another earth

It is useless to activate the jump trick from E5 to E6 because if I do so before OOB I'm in the same "level" than the boulder which is loaded anyway so I can't trigger the maze (as I usually load E6 by under the boulder).

2:08: It is unfortunate but I wont go back to the game for this.

I saw the "freeze" thing too, weird =o you can't even open the menu anymore

tigrou

Edit history:

tigrou: 2016-06-18 04:59:45 am

tigrou: 2016-06-18 04:34:06 am

tigrou: 2016-06-18 04:33:53 am

tigrou: 2016-06-17 02:51:49 pm

tigrou: 2016-06-17 02:12:29 pm

tigrou: 2016-06-17 02:11:50 pm

tigrou: 2016-06-17 02:07:13 pm

tigrou: 2016-06-17 02:05:35 pm

tigrou: 2016-06-17 02:05:11 pm

tigrou: 2016-06-17 02:01:07 pm

tigrou: 2016-06-17 02:00:51 pm

tigrou: 2016-06-17 12:40:10 pm

tigrou: 2016-06-17 12:39:18 pm

tigrou: 2016-06-17 12:38:17 pm

tigrou: 2016-06-17 12:36:35 pm

tigrou: 2016-06-17 12:36:29 pm

tigrou: 2016-06-17 12:36:14 pm

2016-06-17 12:34:53 pm

@NHG : the run is pretty cool. I especially like the moment when Emily land up just in front of pregtz and barely avoid the fireball.
Even if it is probably perfectible, it is better than any previous run on that game, and thus the new world record. Good job.

I have one question : how is the "Emily stays in the air" trick performed ? I mean how to trigger it ?
I know it has been discussed previously but I did not followed everything.

About the run itself and possible lag (as pointed out by LotBlind) : if you use GOG version with default settings (CPU cycles = 11000) the game is clearly not running at full speed.
By experience, while watching the NHG video, I can clearly feel there is parts where the game slow down.

You can verify that yourself by turning on FPS display : change DOSBOX "graphic mode" to "opengl" (there is a shortcut called "DOSBOX configurator" in GOG.com folder) and start FRAPS.
In lot of places, the game will run at 60 fps (what it try to target) but as soon as you will have some enemies on the screen it will go down to around 35 fps. Transitions between cameras are also pretty laggy/slow.

The 11000 CPU cycles is only set by GOG as a suggestion. At that speed, it barely emulate a fast 386 (see here)
From my own experience to have it run at full speed you need at least 20000 CPU cycles (press CTRL-F11/F12 to decrease/increase them).
You can also try to set the parameter "cycles" to "max" in dosboxAlone1.conf file (cycles=max). In that case DOSBOX will use 100% of a CPU core. I tried that on my machine and the game is very smooth. I have zero slowdown and camera/room transitions are blazing fast.

EDIT : before you start doing anything serious with the room viewer, always make sure you have latest version. I fix bugs and add small improvements from time to time.

@NHG : the run is pretty cool. I especially like the moment when Emily land up just in front of pregtz and barely avoid the fireball.
Even if it is probably perfectible, it is better than any previous run on that game, and thus the new world record. Good job.

I have one question : how is the "Emily stays in the air" trick performed ? I mean how to trigger it ?  
I know it has been discussed previously but I did not followed everything.

About the run itself and possible lag (as pointed out by LotBlind) : if you use GOG version with default settings (CPU cycles = 11000) the game is clearly not running at full speed. 
By experience, while watching the NHG video, I can clearly feel there is parts where the game slow down.

You can verify that yourself by turning on FPS display : change DOSBOX "graphic mode" to "opengl" (there is a shortcut called "DOSBOX configurator" in GOG.com folder) and start FRAPS.
In lot of places, the game will run at 60 fps (what it try to target) but as soon as you will have some enemies on the screen it will go down to around 35 fps. Transitions between cameras are also pretty laggy/slow.

The 11000 CPU cycles is only set by GOG as a suggestion. At that speed, it barely emulate a fast 386 (see [url=https://www.dosbox.com/wiki/Performance#Emulated_CPU_equivalency]here[/url])
From my own experience to have it run at full speed you need at least 20000 CPU cycles (press CTRL-F11/F12 to decrease/increase them).
You can also try to set the parameter "cycles" to "max" in dosboxAlone1.conf file (cycles=max). In that case DOSBOX will use 100% of a CPU core. I tried that on my machine and the game is very smooth. I have zero slowdown and camera/room transitions are blazing fast.

EDIT : before you start doing anything serious with the room viewer, always make sure you have latest version. I fix bugs and add small improvements from time to time.

NHG

2016-06-17 02:46:48 pm

Shut up and go back to another earth

To trigger to jump-trick, simply jump and open the item list during the jump. Then drop an item. The action'll be cancelled (Emily won't drop anything) but she'll be "stuck" at the same height until the next jump or climb.

As for the speed of the game, I know it is possible to modify some DOSBOX settings but I think it's not a good thing to change it if I want to submit it on this website. In any case it is not really important as long as every run use the same speed. And to be honest the suggested speed is probably even slightly faster than the original one, you can clearly notice that if you watch the introduction, the music is cut two times wich is normally not the case.

Aqfaq

2016-06-19 11:51:30 am

Hi NHG, your latest run is something only a true speedrunner can do. Really good gaming skills there. It is safe to say that the whole thing is pure art by now. Lovecraft would be proud.

Hi LotBlind, great work with everything, everything and everything as always!

Hi tigrou, the world is a better place because of people like you.

Hi Kelly, hi Kelly!

LotBlind

2016-06-19 01:48:14 pm

tigrou: is there any way we could read the lag from the game's memory? I.e. is there anything that would act as a sort of "time since last frame drawn" counter, or even time since the last input was read, that could be displayed in the viewer itself? I don't/can't use FRAPS, just native DosBox recording. If we could see when the game reads inputs I could at least do something like change the input I'm giving while the lag is happening... come to think of it we can already do that with the room/view transitions because all you need to do is monitor values we already can see: character's position or direction, and see when they've last changed. That doesn't work in the menu though...

Something like that would make it a helluva lot easier to try to optimize the run more for lag reduction. BTW are you guys sure the lag during when you play is the same as what the recording will have? Even if I use DosBox recording?

NHG

2016-06-20 04:01:13 am

Shut up and go back to another earth

Oh, thank you Aqfaq it's really nice ^^ and by the way thank you LotBlind Tigrou and Mr. Kelly too ! How rude I am.

Maybe a good way to test lag effect could be to trigger some ghosts in the ballroom ? I already told you that but DOSBOX seemed to cancel the lag when I recorded a run with it (I touched a ghost who made the game slowdown gradually until the end, yet when I watched the video it was perfectly fine like the lag never existed), it was with another computer though, I never tested with this one.

LotBlind

2016-06-21 08:39:56 am

Now I'm wondering: clearly DosBox ignores its own lag frames, but external software doesn't... That means our timings will be incompatible, seeing as there clearly is lag at 11000 cycles. I wonder if others running Dos games have used external or internal recording. I'll be asking about that next. Because there's no way I can record the run then... unless there's 2 kinds of lag?

The rules page states you're okay to record on DosBox, which means I can do the run, but I still don't get how we can compare run times... I wonder if you should have been using DosBox recording too? It just makes no sense not to. With some games this doesn't properly record the sound but in AitD it obviously does.

Here's the thread where I posed the question about lag: https://forum.speeddemosarchive.com/post/time_for_a_new_update_of_the_sda_rules_66.html#time_for_a_new_update_of_the_sda_rules_66